glusterfs.git/xlators/nfs, branch v3.7dev

make snapview-server more compatible with NFS server

2014-07-16T09:27:50+00:00

* There was no handle based API for listxattr. With this change, glfs_h_getxattrs
  also handles the listxattr functionality by checking whether the name is NULL
  or not (like posix). But all the gfapi functions for listxattr
  (glfs_h_getxattrs AND glfs_listxattr AND glfs_flistxattr) returns the names of
  the xattrs in a buffer provided by the caller. But snapview-server has to
  return the list of xattrs in a dict itself (similar to posix xlator). But
  the buffer just contains the names of the xattrs. So for each xattr, a zero
  byte value is set (i.e. "") into the dict and sent back. Translators which
  do xattr caching (as of now md-cache which caches selinux and acl related
  xattrs) should not cache those xattrs whose value is a zero byte data ("").
  So made changes in md-cache to ignore zero byte values.

* NFS server was not linking the inodes to inode table in readdirp. This was
  leading to applications getting errors. The below set of operations would
  lead to applications getting error
  1) ls -l in one of the snaopshots (snapview-server would generate gfids for
     each entry on the fly and link the inodes associated with those entries)
  2) NFS server upon getting readdirp reply would not link the inodes of the
     entries. But it used to generate filehandles for each entry and associate
     the gfid of that entry with the filehandle and send it as part of the
     reply to nfs client.
  3) NFS client would send the filehandle of one of those entries when some
     activity is done on it.
  4) NFS server would not be able to find the inode for the gfid present in the
     filehandle (as the inode was not linked) and would go for hard resolution
     by sending a lookup on the gfid by creating a new inode.
  5) snapview-client will not able to identify whether the inode is a real inode
     existing in the main volume or a virtual inode existing in the snapshots
     as there would not be any inode context.
  6) Since the gfid upon which lookup is sent is a virtual gfid which is not
     present in the disk, lookup would fail and the application would get an
     error.

  To handle above situation, now nfs server also does inode linking in readdirp.

Change-Id: Ibb191408347b6b5f21cff72319ccee619ea77bcd
BUG: 1115949
Signed-off-by: Raghavendra Bhat 
Reviewed-on: http://review.gluster.org/8230
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos 
Reviewed-by: Raghavendra G 
Reviewed-by: Vijay Bellur

nfs: prevent assertion error with MOUNT over UDP

2014-07-07T17:28:47+00:00

The MOUNT service over UDP runs in a separate thread. This thread does
not have the correct *THIS xlator set. *THIS points to the global (base)
xlator structure, but GF_CALLOC() requires it to be the NFS-xlator so
that assertions can get validated correctly.

This is solved by passing the NFS-xlator to the pthread function, and
setting the *THIS pointer explicitly in the new thread.

It seems that on occasion (needs further investigation) MOUNT over UDP
does not unregister itself. There can also be issues when the kernel NLM
implementation has been registered at portmap/rpcbind, so adding some
unregister procedures in the cleanup of the test-cases.

Change-Id: I3be5a420fc800bbcc14198d0b6faf4cf2c7300b1
BUG: 1116503
Signed-off-by: Niels de Vos 
Reviewed-on: http://review.gluster.org/8241
Reviewed-by: Santosh Pradhan 
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

gNFS: Fix multi-homed m/c issue in NFS subdir auth

2014-06-25T07:41:52+00:00

NFS subdir authentication doesn't correctly handle multi-homed
(host with multiple NIC having multiple IP addr) OR multi-protocol
(IPv4 and IPv6) network addresses.

When user/admin sets HOSTNAME in gluster CLI for NFS subdir auth,
mnt3_verify_auth() routine does not iterate over all the resolved
n/w addrs returned by getaddrinfo() n/w API. Instead, it just tests
with the one returned first.

1. Iterate over all the n/w addrs (linked list) returned by getaddrinfo().
2. Move the n/w mask calculation part to mnt3_export_fill_hostspec()
   instead of doing it in mnt3_verify_auth() i.e. calculating for each
   mount request. It does not change for MOUNT req.
3. Integrate "subnet support code rpc-auth.addr..allow"
   and "NFS subdir auth code" to remove code duplication.

Change-Id: I26b0def52c22cda35ca11766afca3df5fd4360bf
BUG: 1102293
Signed-off-by: Santosh Kumar Pradhan 
Reviewed-on: http://review.gluster.org/8048
Reviewed-by: Rajesh Joseph 
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos

gNFS: Make NFS DRC off by default

2014-06-09T07:27:07+00:00

DRC in NFS causes memory bloat and there are known memory corruptions.
It would be good to disable drc by default till the feature is stable.

Change-Id: I93db6ef5298672c56fb117370bb582a5e5550b17
BUG: 1105524
Original-patch-by: Santosh Kumar Pradhan 
Signed-off-by: Niels de Vos 
Reviewed-on: http://review.gluster.org/8004
Tested-by: Gluster Build System 
Reviewed-by: Krishnan Parthasarathi 
Reviewed-by: Santosh Pradhan 
Reviewed-by: Vijay Bellur

rpcsvc: Validate RPC procedure number before fetch

2014-05-17T18:56:01+00:00

While accessing the procedures of given RPC program in,
rpcsvc_get_program_vector_sizer(), It was not checking boundary
conditions which would cause buffer overflow and subsequently SEGV.

Make sure rpcsvc_actor_t arrays have numactors number of actors.

FIX:
Validate the RPC procedure number before fetching the actor.

Special Thanks to: Murray Ketchion, Grant Byers

Change-Id: I8b5abd406d47fab8fca65b3beb73cdfe8cd85b72
BUG: 1096020
Signed-off-by: Santosh Kumar Pradhan 
Reviewed-on: http://review.gluster.org/7726
Tested-by: Gluster Build System 
Reviewed-by: Rajesh Joseph 
Reviewed-by: Anand Avati

mgmt/gluster: Use fsync instead of O_SYNC

2014-05-05T20:52:51+00:00

Glusterd uses O_SYNC to write to temp file then performs renames
to the actual file and performs fsync on parent directory. Until
this rename happens syncing writes to the file can be deferred.
In this patch O_SYNC open of temp file is removed and fsync of the
fd before rename is done.

Change-Id: Ie7da161b0daec845c7dcfab4154cc45c2f49d825
BUG: 908277
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/7370
Reviewed-by: Krishnan Parthasarathi 
Reviewed-by: Raghavendra Bhat 
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

protocol/server: Validate outstanding-rpc-limit

2014-05-03T19:43:25+00:00

server.outstanding-rpc-limit by passes validation enforcement of gluster
as protocol/server ignores any keys starting with rpc (rpc* wildcard).
Which also causes a side effect, i.e. help text gets displayed as NULL.

Change-Id: I925e028529d671e1e29a0c7ff7e625fdb49a1be4
BUG: 1091364
Signed-off-by: Santosh Kumar Pradhan 
Reviewed-on: http://review.gluster.org/7562
Reviewed-by: Poornima G 
Reviewed-by: Krishnan Parthasarathi 
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

cli/hooks : Add volume set options to enable/disable nfs-ganesha support.

2014-05-03T14:24:07+00:00

1. gluster volume set nfs-ganesha.enable ON/OFF
If the option is set to ON, the volume field in the nfs-ganesha configuartion file is
edited. Gluster-nfs is disabled on that volume and the volume is exported using
nfs-ganesha.

2.gluster volume set nfs-ganesha.host IP
This is used to provide the IP of the nfs-ganesha host.

Note : nfs-ganesha.host MUST be set before using nfs-ganesha.enable ON

The switch from gluster-nfs to nfs-ganesha is mostly done by the hook-scripts
in the post phase of the 'set' option. As a result, gluster volume reset does not
function as it is expected to. By default, nfs-ganesha will be set to off but the
process  will not be killed.

Hence, a few changes have to be made post 'reset' option as well. Those changes
also have been added.

Change-Id: I7fdc14ee49d1724af96eda33c6a3ec08b1020788
BUG: 1092283
Signed-off-by: Meghana 
Reviewed-on: http://review.gluster.org/7321
Reviewed-by: Raghavendra Talur 
Reviewed-by: Santosh Pradhan 
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

nlm: fix rpc.statd typo

2014-04-29T01:24:22+00:00

Change-Id: I701fd82a8dd5a72727b8035bc6c2861465f1aa1f
BUG: 1089172
Signed-off-by: Harshavardhana 
Reviewed-on: http://review.gluster.org/7585
Reviewed-by: Anand Avati 
Tested-by: Anand Avati

gNFS: Log properly for pmap_getport() fail in NLM

2014-04-27T03:00:55+00:00

In NLM callback path, if pmap_getport() FAILs, it just log the error
message saying "Is firewall running on the client". It may happen
that RPC services are not running e.g. "rpcbind" is not running or
nlockmgr (NLM) is not registered with portmapper which all can be
checked using "rpcinfo -p" command.

FIX:
Modify the log message to include the later case mentioned
above.

Change-Id: If422275b2ab59d1e974a6caa37132f31e9a34329
BUG: 1090782
Signed-off-by: Santosh Kumar Pradhan 
Reviewed-on: http://review.gluster.org/7544
Reviewed-by: Niels de Vos 
Tested-by: Gluster Build System 
Reviewed-by: Rajesh Joseph